Algorithm for synthesis of a neuro-fuzzy discrete regulator in tasks of control of a dynamic object
- № 4 (52) 2019
Страницы:
44
–
52
Язык: английский
Аннотация
В статье приводятся результаты исследования свойств систем автоматического регулирования с пропорционально-интегрально- дифференциальном-регулятором и последовательным адаптивным корректирующим устройством динамических характеристик систем автоматического регулирования. Для настройки параметров пропорционально-интегрально-дифференциального регулятора предложена двухслойная искусственная нейронная сеть, отличающаяся простотой реализации на микроконтроллерах. Для придания свойства адаптации цифровому пропорционально-интегрально-дифференциальному регулятору использован метод интерактивной адаптации, основанный на методе Фреше.Приведены результаты исследования свойств системы автоматического регулирования с пропорционально-интегрально-дифференциальным регулятором и последовательным адаптивным корректирующим устройством динамических характеристик систем автоматического регулирования.
Мацолада пропорционал-интеграл-дифференциал-регулятор ва автоматик бошцариш тизимларининг динамик характеристикаларини кетма- кет мослаштирадиган динамик автоматик бошцарув тизимининг хусусиятларини урганиш натижалари берилган. Пропорционал-интеграл- дифференциал-регулятор параметрларини созлаш учун микроконтроллерларда амалга оширишнинг соддалиги билан ажралиб турадиган икки цаватли сунъий нейрон тармоц таклиф этилади. Мослашув хусусиятини рацамли пропорционал-интеграл-дифференциал-регуляторга бериш учун Фреше усулига асосланган интерфаол мослашув усули цулланилади. ПИД-регулятор ва автоматик бошцарув тизимларининг динамик характеристикаларини кетма-кет мослаштирувчи тузатувчи мосламали автоматик бошцарув тизимининг хусусиятларини урганиш натижалари берилган.
The article presents the results of a study of the properties of an automatic control system with a proportional-integral-differential-controller and a sequential adaptive corrective device for the dynamic characteristics of automatic control systems. To adjust the parameters of the proportional-integral-differentialcontroller, a two-layer artificial neural network is proposed, which is distinguished by the simplicity of implementation on microcontrollers. To give the adaptation property to the digital proportional-integral-differential-controller, the method of interactive adaptation based on the Frechet method is used. The results of the study of the properties of an automatic control system with a proportional-integraldifferential-controller and a serial adaptive corrective device for the dynamic characteristics of automatic control systems are presented. A distinctive feature of adaptation lies in the fact that during the operation of the system the regulator parameters do not change and correspond to the setting preceding the start of the system. In this case, depending on the change in the parameters of the control object, the transfer coefficient of the corrector or the phase shift created by it changes. Changes that occur only in cases where the quality of the automatic control system becomes unsatisfactory due to changes in the properties of the control object or due to the impact of disturbances on the control object. To give the desired property to control systems, a non-linearity of the “saturation” type is included in the control loop. The use of nonlinear correctors is associated with the problem of taking into account the dependence of the frequency characteristics on the amplitude of harmonic oscillations of the input signal, which allows one to obtain the required amplitude and phase frequency characteristics, which ensures the stability of the system and improves the quality of the transition process. A comparative analysis of the results obtained by conducting a simulation experiment on the environment «MATLAB». The results of the study, an automatic control system with linear corrective devices showed that the adaptation of linear correctors even using modern mathematical apparatus, such as fuzzy logic, is possible only with limited range and nature of changes in the parameters of the control object. The results can be implemented on a production system, i.e. to maintain the given mode of the boiler units in the power system.
Мацолада пропорционал-интеграл-дифференциал-регулятор ва автоматик бошцариш тизимларининг динамик характеристикаларини кетма- кет мослаштирадиган динамик автоматик бошцарув тизимининг хусусиятларини урганиш натижалари берилган. Пропорционал-интеграл- дифференциал-регулятор параметрларини созлаш учун микроконтроллерларда амалга оширишнинг соддалиги билан ажралиб турадиган икки цаватли сунъий нейрон тармоц таклиф этилади. Мослашув хусусиятини рацамли пропорционал-интеграл-дифференциал-регуляторга бериш учун Фреше усулига асосланган интерфаол мослашув усули цулланилади. ПИД-регулятор ва автоматик бошцарув тизимларининг динамик характеристикаларини кетма-кет мослаштирувчи тузатувчи мосламали автоматик бошцарув тизимининг хусусиятларини урганиш натижалари берилган.
The article presents the results of a study of the properties of an automatic control system with a proportional-integral-differential-controller and a sequential adaptive corrective device for the dynamic characteristics of automatic control systems. To adjust the parameters of the proportional-integral-differentialcontroller, a two-layer artificial neural network is proposed, which is distinguished by the simplicity of implementation on microcontrollers. To give the adaptation property to the digital proportional-integral-differential-controller, the method of interactive adaptation based on the Frechet method is used. The results of the study of the properties of an automatic control system with a proportional-integraldifferential-controller and a serial adaptive corrective device for the dynamic characteristics of automatic control systems are presented. A distinctive feature of adaptation lies in the fact that during the operation of the system the regulator parameters do not change and correspond to the setting preceding the start of the system. In this case, depending on the change in the parameters of the control object, the transfer coefficient of the corrector or the phase shift created by it changes. Changes that occur only in cases where the quality of the automatic control system becomes unsatisfactory due to changes in the properties of the control object or due to the impact of disturbances on the control object. To give the desired property to control systems, a non-linearity of the “saturation” type is included in the control loop. The use of nonlinear correctors is associated with the problem of taking into account the dependence of the frequency characteristics on the amplitude of harmonic oscillations of the input signal, which allows one to obtain the required amplitude and phase frequency characteristics, which ensures the stability of the system and improves the quality of the transition process. A comparative analysis of the results obtained by conducting a simulation experiment on the environment «MATLAB». The results of the study, an automatic control system with linear corrective devices showed that the adaptation of linear correctors even using modern mathematical apparatus, such as fuzzy logic, is possible only with limited range and nature of changes in the parameters of the control object. The results can be implemented on a production system, i.e. to maintain the given mode of the boiler units in the power system.